Identification of disinfection by-product precursors by natural organic matter fractionation: a review

During disinfection of drinking water, natural organic matter reacts with chlorine to produce harmful disinfection by-products. The identification of precursors of disinfection by-products in natural organic matter is challenging because natural organic matter is very complex and poorly known. Therefore, scientists have focused on the fractionation of natural organic matter with membranes or resins to better understand how and which organic matter fractions react during chlorination. Here, we compared the reactivity of various organic fractions with disinfection by-products. For that we did a meta-analysis of 400 water samples published in 80 publications, with focus on chlorination time and dose, SUVA254 and the column capacity factor used during resin fractionation. SUVA254 refers to the ultraviolet absorbance at 254 nm divided by the organic matter concentration. We found that hydrophobic compounds have 10–20% higher reactivity to both trihalomethane and haloacetic acid formation compared to hydrophilic compounds in waters with SUVA254 above 2L/(mg∙m), while hydrophobic and hydrophilic compounds have equal reactivity in waters with low SUVA254. On the other hand, hydrophilic compounds are 20–80% more reactive towards emerging disinfection by-products, regardless of SUVA254. Chlorination time and dose do not influence the reactivity ratio between the different fractions. An increase in column capacity factor can shift the reactivity ratio from hydrophobic to hydrophilic fractions. Dead-end, stirred cell ultrafiltration membrane fractionation might not always produce sharply separated fractions, which is mainly due to fouling. Therefore, no clear correlation could be found between membrane fractions and all investigated disinfection by-product groups.